Paraneoplastic neurological syndromes (PNSs) occur in the instance of a cancer, often before its discovery, and are not connected to either the tumor proliferation itself (direct invasion, metastases) or the treatment. Their frequency is estimated to be, overall, approximately 1% of cancers. Several clinical pictures have been individualized for a long time (encephalomyelitis, Denny-Brown sensory neuropathy, cerebella atrophy, limbic encephalitis, opsoclonus, etc.) corresponding in fact to the either elective or preferential attack of certain groups of neurons. The frequency of inflammatory cells in the vicinity of the lesions for many years brought to mind the possibility of an autoimmune or viral process. The more recent demonstration of autoantibodies in the serum and in the cerebrospinal fluid (CSF) of patients suffering from PNS, specific for the type of tumor and for the type of neurons which degenerate, has revived the hypothesis that autoimmunity contributes to generating this pathological condition (Graus et al., 1985; Greenlee et al., 1983).
Besides the presence of a high titer of these antibodies in the blood and the CSF of patients, there are several arguments suggesting that PNSs are the product of autoimmune mechanisms. Thus, the antigens recognized in the central nervous system are also present in the tumors of patients (Anderson et al., 1987). Antibodies specifically directed against these antigens and also B and T lymphocytes are found within the tumor tissue (Hetzel et al., 1990).
These data suggest that the autoimmune process could be triggered by the expression of tumor antigens. A crossed immunity process could cause the lesions in the central nervous system. Other arguments also indicate that the cerebral lesions result from the autoimmune response. Thus, in the brain of the patients, the specific antibody titer is higher than that of the serum and of the CSF (Dalmau et al., 1991). In addition, in the case of encephalomyelitis associated with anti-Hu antibodies, there is an intense lymphocytic reaction, made up of B and T cells, located in proximity to neurons undergoing destruction (Dalmau et al., 1991; Graus et al., 1990).
Several types of autoantibody allowing precise syndromic groupings as a function of immunological, neurological and cancer-related criteria have been described.
Thus, anti-Yo antibodies are found in the serum and the CSF of women having paraneoplastic cerebella atrophy and a gynecological cancer (ovary, breast or uterus) (Greenlee et al., 1983; Jaeckle et al., 1985).
These antibodies recognize two cytoplasmic proteins, of 34 and 62 kDa, specific for Purkinje cells of the cerebellum.
Anti-Ri antibodies are found in the serum and the CSF of patients (mainly women) having opsomyoclonus, a cerebella syndrome and breast cancer. These antibodies recognize two proteins, of 50 and 80 kDa, specific for central nervous system neurons (Luque et al., 1991).
Anti-Hu antibodies are most commonly encountered in the course of PNSs. They are found in the serum and the CSF of patients having Denny-Brown's syndrome or encephalomyeloneuritis and small-cell lung cancer (Graus et al., 1985; Dalmau et al., 1992). These autoantibodies recognize several proteins, of 37 to 45 kDa, expressed specifically by all the neurons of the nervous system.
Another type of autoantibody has been identified in patients having PNS: anti-CV2 antibodies (Antoine et al., 1993; Honnorat et al., 1996). The latter are atypical, in the sense that the antigenic target recognized in adulthood is essentially non-neuronal, although post-mortem analysis of the brain of four patients makes it possible to observe neuronal loss, gliosis and an inflammatory process characteristic of PNSs.
The originality of the discovery of these autoantibodies lies, firstly, in their demonstration. The latter had evaded all the usual investigations which consisted in revealing the antigens recognized, by immunohistochemistry on post-mortem brain. The antigen recognized is in fact soluble and disappears from post-mortem brain under the majority of conditions for fixing. Only fixing of human post-mortem tissue by immersion in paraformaldehyde, or fixing in situ by perfusion of paraformaldehyde in animals, has made it possible to reveal the presence of these antibodies in the CSF or the serum of patients suffering from PNS (Antoine et al., 1993; Honnorat et al., 1996).
The anti-CV2 autoantibodies present in the sera of patients suffering from paraneoplastic neurological syndrome (PNS) have been defined by their ability to recognize, by indirect immunohistochemistry, a cytoplasmic antigen expressed specifically, in adult rat brain, by a subpopulation of oligodendrocytes of the brain stem, of the medulla and of the cerebellum.
The originality of these autoantibodies lies, secondly, in their diagnostic value. Their presence in the serum or the CSF of patients is of diagnostic value since it makes it possible to specify the paraneoplastic origin of a neurological syndrome. The discovery of these antibodies, when it precedes that of cancer, directs the search for this cancer and enables it to be discovered. Such was the case for six patients out of 19 having anti-CV2 antibodies. The clinical disorders were different depending on the patients, some of them exhibiting a picture of limbic encephalitis, others encephalomyeloneuritis and others Lambert-Eaton syndrome. Nevertheless, in more than 60% of cases, the cerebella syndrome was predominant. The most commonly associated tumor was small-cell lung cancer (60% of cases).
Experiments on newborn rat brains have shown that these anti-CV2 antibodies react with a 66 kDa protein (Honnorat et al., 1996).
In adult brain, this antigen is located in a subpopulation of oligodendrocytes or in cells which retain differentiation capacities in the adult brain (olfactory bulb, dentate gyrus). The antigen recognized is thought to play a role in neuronal survival, via neuron/oligodendrocyte interactions, as suggested by the loss of neurons observed in the post-mortem brain of patients suffering from PNS.
Its very restricted expression in adulthood contrasts with very strong and transient expression in the central and peripheral nervous system in development, suggesting the probable role of this antigen in the development of the nervous system.
In application WO 98/37192 and the article by Honnorat et al. from 1999, the target antigen of the anti-CV2 autoantibodies, which corresponds to a protein designated “POP-66” for “paraneoplastic oligodendrocyte protein 66 kDa”, was identified as being the human form of the ULIP-4 protein. The ULIP (for “Unc-33 like phosphoprotein”) proteins are involved in the control of neuronal development and axonal transport (Byk et al., 1996). Four members of this family had been identified by three different teams (Byk et al., 1998, Wang and Strittmatter, 1996; and Hamajima et al., 1996). A thorough search for possible other members of this family had come to nothing.
The authors of the present invention were then confronted with new results, which were not coherent with the identification established by application WO 98/37192; although all the anti-CV2 sera tested on Hela cells unquestionably recognized the recombinant ULIP 4 protein in immunohistochemistry, only 20% of these sera recognized the ULIP 4 protein by Western blotting on these same cell extracts. Now, all the anti-CV2 sera tested recognized, moreover, by Western blotting, the same 66 kDa protein after immunoprecipitation, on brain extracts. In addition, the ULIP4 mRNA was only very weakly expressed in oligodendrocytes by in situ hybridization. Based on these data, the authors of the present invention supposed that a new member of the ULIP family, unidentified to date, strongly expressed by oligodendrocytes and recognized in Western blotting by all the anti-CV2 sera, could exist.